PhD position in geochemistry/geotechnic

Title :

Experimental and numerical study of the durability of geomaterials by coupled Thermo- Hydro-Mechanical and Geochemical (THM+geoC) approach applied to geotechnical structures.

Objectives
The aim of the thesis is to lay the foundations for modeling the processes that cause the physico-chemical degradation of materials and their impact on the mechanical stability and hydraulic properties of structures. This PhD thesis aims to couple geochemistry and geotechnic. The aim is to develop an approach that is both numerical and experimental in order to couple the degradation of materials in their environment (geochemical approach) and the consequences of these degradations from the geotechnical point of view (loss of performance). It is clear that few studies link these two aspects, which are generally treated separately by two research communities. Indeed, much research has been conducted on the transport of pollutants in porous materials in relation to environmental problems. But the subject of reactive fluid transport, in a context of durability of earthen structures (for example embankments, dikes, grounds for foundations ...) has not been much developed, except in the particular context of the geological disposal of radioactive waste. In this case, the fluids come into physico-chemical interaction with the porous matrix and cause a modification of the mechanical and hydraulic properties which impacts the service life or the integrity of the structure and can generate a risk situation (rupture, collapse…).

Case study

This study will focus on two case studies: on the one hand, the development of sinkholes in limestone rocks that can lead to the collapse of cavity roofs (ISSMGE, 2006) and on the other hand, the degradation of lime treated soils (even cement treated soils) which are used as materials for protective dykes in fluvial or maritime environment. In both cases, the conditions to which the materials are subjected, especially the contact with water and air, cause their slow and progressive degradation via dissolution and / or precipitation processes coupled with the transport of the chemical species in the material, which leads to the creation of a chemical gradient and finally to a loss of material when the species leaves the material

Experimental approach

The approaches currently used to study alteration and/or durability of geomaterials are based on laboratory tests carried out on multi-centimeter specimens such as accelerated degradation tests (eg. immersion cycles in hot water and then a drying in oven). The test is followed by a measurement, usually a measure of the mechanical strength. The approach consists in observing macroscopically a overall effect, that is to say the decrease of the material performance, when it is subjected to an aggressive environment. We propose to go further during this PhD by introducing the physico-chemical reactions involved in the material (dissolution / precipitation, adsorption, oxidation-reduction ...) and the transport of dissolved substances in the material (kinetic effect and formation of gradient). Added to this, there is the possibility of material transport (movement of fine particles) via percolating fluids or the possible degradation via bacterial or fungal activity, for example.

An approach based on the acquisition of percolation data coupled with geochemical-transport modeling is proposed. The percolation tests consist in passing a fluid of known initial composition through a macroscopic geomaterial specimen (h = 10 cm and f  = 5 cm). The overall response of the system will be analyzed from the measurement of percolated volumes (relative to the pore volume of the specimen), percolation time and chemical composition of the percolated water (average measurements by ICP/OES and/or with continuous measurement with sensors positioned at the outlet). The breakthrough curves of the chemical elements of interest can thus be determined.

To evaluate the internal degradation of the specimens and the gradients that take place between the upstream and the downstream, it is proposed to separately study different slices cut of the tested specimen after a time t (variable) of damage by percolation. It will be measured the evolution of the microstructure (by SEM coupled to EDX to perform a mapping followed by image processing, mercury intrusion porosimetry or even X-ray tomography, petrographic approach by optical microscopy) and quantitative mineralogical characterization (by XRD, ATD-ATG). This experimental approach at the micro scale will be compared with the degradation of the mechanical and hydraulic properties of the specimens having undergone the percolation step: nondestructive pund it test (or speed of sound) and water permeability test), measure with a triaxial cell or rock permeability bench, simple compression test, dynamic Young's modulus measurement... As far as possible these properties should be measured on each specimen slice (if the test configuration allows it). A micro-indentation approach should be explored to obtain locally the mechanical properties of the material.

Numerical approach

The innovative aspect in our approach is the coupling between the experimental study and the numerical modeling of both chemical and mechanical processes using the following codes:

• The PhreeqC-3 code (Parkhust and Appelo) used widely internationally is a geochemical calculation code that can take into account a complex reactive system but with a 1D water transport module whose performance is limited. This model calculates the equilibrium concentrations of the chemical species present in a water-solid-gas system, for example calcite in the presence of a solution of NaCl at pH 3 will dissolve to give calcium and carbonate ions. The kinetics of the processes remains however a subject in development in this computational code. In another case, if we put certain amount of reactive substances in contact with a matrix, the code quantifies pollutants that remain in solution after adsorption by the solid. Note that geochemical calculations are highly dependent on the chemical reactions and their constants that are introduced into the database associated with the code. If necessary, we will add the chemical reactions and their constants which would be missing in relation to the geomaterial, by performing batch experiments using model systems (study of interactions in separated reactor). Otherwise, these parameters will be considered as variables, and in this case, we will estimate their value and their impact on the final result of the geochemical models then that of the complete calculations of reactive transport.
• the Disroc calculation code (Fracsima, 2016). Generally, the evolution of the microstructure is little or not taken into account in the modeling when for example the material dissolves during the percolation. The Disroc code has the possibility of modeling the coupling between Thermo-Hydro-Mechanical phenomena (including water transport) and chemical processes limited to a preponderant reaction. Our objective will therefore be to use this possibility to not only predict reactive transport (coupling of geochemistry and water transport) but also to model the effect of chemical alteration on the mechanical strength or hydraulic integrity of the geomaterial. This important step to understand the evolution of a structure or site is offered by the Disroc code subject to certain adjustments according to our needs (the code is sufficiently open to allow the user to introduce his own chemical reaction laws or chemistry-transport or chemical-mechanical coupling).

Finally, the goal will be to find how to link these two computational codes that couple the reactivity of the material, the transport of the dissolved species in the matrix (gradient effect) and finally the effects on use properties (mechanical and hydraulic). The results obtained numerically will first be compared with the experimental tests for percolation and characterization of post-percolation specimens in order to identify the parameters of the laws by an inverse analysis. Numerical modeling will then be applied to the two simple case studies defined above: the phenomenon of sinkholes (fontis) creation and the durability of dikes. More precisely:

• To test the modeling tools and the experimental approach, the study of the effects of percolation of water in a homogeneous calcareous material (porous medium supposed to be 'simple' and representative of a karst environment where cavities can develop) will serve as a starting point. The effect of an acid disturbance (percolation of an acidified solution via CO₂ in the air) will be simulated. We will study the effect of water composition (salinity), pH or even P_CO2 (calco-carbon equilibria), the Ca2 + content of percolating water and the presence of other ions (inhibitory effect or activator of the dissolution) or the presence of NaCl salt (or seawater)... The degree of saturation of the material will be maintained at 100% with a continuous percolation with renewed water (we will not consider the effects of desaturation or drying). The effect of these degradations on the mechanical stability of a sinkholes (fontis) structure will be modeled numerically.
• Our approach will be applied to a second experimental system, that is to say a lime-treated soil that can be found in river or sea dikes. The mixtures studied will be for example a compacted mixture of lime and clay. The clays in contact with hydrated lime or portlandite allow the development of slow pozzolanic reactions at the origin of the formation of cement hydrates. However, the performance of this initially resistant mixture is reduced over time by leaching by rainwater (percolation). Experimental tests must confirm the processes causing the degradation of the material and the effects of this degradation. We will then try to approach complex natural systems that are multiphasic (complex mineralogy and microstructure) by introducing sand and silt into the mixture. The more complex the system is, the more it will be necessary to make assumptions to simplify it. The evaluation of preponderant reactions is a key step in achieving our goal and is the first step in geochemical modeling. Finally, the effect of the modifications of the mechanical and hydraulic properties resulting from these reactions on the stability of a dike structure will be analyzed by numerical modeling.

For more information go to :https://www.ifsttar.fr/offres-theses/sujet.php?num=2196&num_session=1&ver=an

The site to candidate  : https://www.ifsttar.fr/offres-theses/index_en.php